Most building and trades businesses don't have an AI problem. They have a handover problem. A job gets quoted in a quoting tool, scheduled in another, priced again by a supplier, photographed on site, and invoiced from whichever version someone finds first. AI can help with that, but only once you know where the handovers are and who checks them.
This checklist is for owners and operations leads in construction and the trades who are being pitched AI and want a plain way to work out whether the business is ready, and where to start. Print it, work through it with the person who actually does the admin, and tick what's true today.
The work: do you know where the hours go?
- We can walk a recent job from first enquiry to paid invoice and list every time someone retyped, checked or chased something.
- We know which handovers cause the most rework (for example quote to job, site report to office, supplier invoice to job cost, or variation to claim).
- We have written that list down with the people who do the work, not from memory.
- We can say roughly how often each of those handovers happens in a normal week.
If you can't tick these yet, start here. The list is worth more than any demo, and it gives you something concrete to hold a seller to.
Your systems: what do you already run?
- We have a list of every tool the business pays for: estimating, job management, scheduling, accounting, forms, file storage and messaging.
- We know which of those tools people actually use day to day, and which ones we pay for but only use a small slice of.
- We know where the "source of truth" is for jobs, clients, prices and costs, or we know that it's currently split across several places.
- We have checked whether the tools we already own can do some of this work with settings we haven't switched on.
Plenty of businesses find they don't have a software shortage. They have plenty of tools that don't talk to each other. Often the most useful first step is connecting what you've got, not adding another login.
Your data: is it in a state a system can use?
- Job numbers or job names follow a pattern the team sticks to.
- Cost codes, stages or categories mean the same thing in the estimating tool, the job system and the accounts.
- Site photos, forms and documents are stored against the job, not scattered across personal phones and inboxes.
- We know who is allowed to change prices, rates and templates, and where the current version lives.
None of this has to be perfect. But if different people would code the same invoice differently, an automated system will inherit that confusion. Tidying the basics is often part of the first project.
Checks and accountability: who signs off?
- For each handover on our list, we know who checks the result before it goes to a client, a supplier or the accounts.
- We are clear that nothing gets paid, lodged or sent to a client without a person approving it, unless we have deliberately decided otherwise.
- We know what a mistake would cost at each step: an annoyed client, a delayed claim, a wrong payment.
- We have a way to see exceptions (the things that didn't match) in a single place, instead of discovering them at month end.
Security and access: what will it touch?
- We know which systems an AI tool would need to read from or write to, and which ones it should never touch.
- We know where client, supplier and staff data would be stored, and who could see it.
- We would not hand over a shared admin password. Access would be limited to what the job needs.
- We have asked any seller what happens to our data if we stop using them.
At AI events, the conversation is usually about what the tools can do. Very few people ask about security. Ask anyway. If keeping things locked down makes the system a fraction slower, that is usually a trade worth making.
Your people: will the team use it?
- The people who do the admin today are involved in deciding what changes.
- We have someone in the business who will own the new way of working after it's built.
- We are clear with the team that the aim is to take the repetitive work off them so they can handle more jobs, not to replace them.
- We have thought about site crews too: anything they need to do on a phone has to be quicker than what they do now.
Tools the team helps shape get used. Tools handed over with "good luck" quietly get ignored.
Choosing who to work with
- Any seller can point at our handover list and say which steps they would take off us, and how.
- They have shown it working on a real job of ours, including the messy parts, not only on sample data.
- They can tell us in a sentence how errors get caught before anything goes out.
- They build on the systems we already run, unless there is a clear reason to replace something.
- We know who owns what gets built, and what support looks like after it goes live.
- We know who to call when it breaks, and it's a person, not only a portal.
If a seller's answer to every question is "move everything onto our platform", or "it doesn't make mistakes", keep looking. Our guide on how to choose an AI consultant in Australia goes deeper on these questions.
Reading your results
Mostly ticks under the work, your systems, your data and checks: you're ready for a small, well-defined first project. Pick the handover that causes the most rework and fix just that. Run it alongside the old way for a few weeks before you trust it.
Gaps under the work or your systems: start with the mapping. Walking a job end to end and listing your tools is free, and it usually shows where the real problem is.
Gaps under your data: plan some tidy-up as part of the first project, not as a separate exercise that never finishes.
Gaps under security or your people: sort these out before anyone connects a tool to your systems. They are the most common reason a technically sound project doesn't stick.
Where to start
Pick a single handover. Fix it with the people who do it. Check it against the old way. Then move to the next. Be wary of anyone who wants to automate the whole business at once: change too much together and a fix in a single place can break things elsewhere.
If you'd like a second pair of eyes on your list, our AI for construction businesses page explains how we work with building and trades businesses, and our guide to accounts payable automation for builders walks through a common handover in detail. If your list raises a bigger question about where AI belongs in the business, our independent AI advisory starts there. When you know which job to fix, our AI implementation service designs, builds and hands over the workflow with your team.
Common questions from construction and trades owners
Can AI help with quotes and tenders in a construction business?
Yes, as a draft that a person checks, not as a finished tender. AI can read a pack, pull out inclusions, exclusions and dates, and turn an estimate into a first draft of a proposal. Your estimator or manager still reviews it before it goes to a client. We would start with the quoting or tender workflow costing you the most hours, alongside the tools you already run.
How do I stop keying supplier invoices into Xero by hand?
Start with the invoice matching your existing software already offers, then connect what it cannot do. A careful design reads the invoice, suggests the job and purchase-order match, and holds anything unusual for a person to approve before it is posted. It should never pay a supplier on its own. There is a longer explainer in our accounts payable article for builders.